Feb. 5, 1956
569
BROMODECARBOXYLATION OF BROMOHYDROXYBENZOIC ACIDS
ORGANIC A N D BIOLOGICAL CHEMISTRY [CONTRIBUTIONFROM
THE
SCHOOL OF CHEMISTRY, GEORGIA INSTITUTE O F
TECHNOLOGY]
Decarboxylation. 111. The Kinetics and Mechanism of Bromodecarboxylation of 3,5-Dibromo-2-hydroxy- and 3,5-Dibromo-4-hydroxybenzoicAcids l v 2
BY ERLINGGROVENSTEIN, JR.,
AND
U. V. HENDERSON, JR.
RECEIVED AUGUST1, 1955 The kinetics of the bromination of 3,5-dibromo-2-hydroxy- and 3,5-dibromo-4-hydroxybenzoicacids and of the corresponding phenols, 2,4-dibromo- and 2,6-dibromophenol, have been studied in 70 to 80% acetic acid a t 20’. At constant hydrogen ion and bromide ion concentrations, these compounds undergo bromination to give tribromophenol a t a rate proportional to the stoichiometric concentration of the reactants. The apparent second-order rate constants ( kapp,) so obtained, however, vary widely with hydrogen ion and bromide ion concentration. After correction for conversion by the bromide ion of much of the bromine to unreactive tribromide ion, the second-order rate constants, k * , still diminish a t constant ionic strength with increasing bromide ion concentration. This decrease is almost negligible for 2,fi-dibromophenol. For 3,5-dibromo-4-hydroxybenzoic acid, on the other hand, k * is almost inversely proportional to the bromide ion concentration a t bromide ion concentrations of 0.04 M or greater. The bromide ion dependency of 3,5-dibromo-2-hydroxybenzoic acid falls in between these two extremes. These bromide ion, as well as the observed hydrogen ion, dependencies are explained on the basis of a mechanism involving attack of molecular bromine to give a reactive cyclohexadienone intermediate which is reverted in part to starting compound by bromide ion or loses carbon dioxide (or hydrogen ion in the case of the dibromophenols) to give tribromophenol. The reversion of the intermediate cyclohexadienone to starting compound is most evident acid and is slight or even negligible for 2,6-dibromophenol. I t is pointed out that similar for 3,5-dibromo-4-hydroxybenzoic mechanisms can explain the kinetics of the iodination of phenol and aniline at iodide ion concentrations of 0.1 M or larger and thus that the effective iodinatim aeent here mav be molecular iodine instead of those previously postulated (I+, HsOI +, IOAc, etc.).
- -
bromine to give tribromophenol, which finally may take up a fourth bromine atom to yield “tribromophenol bromide.” Carbon dioxide is liberated quantitatively0 during the bromodecarboxylation in water. +-Hydroxybenzoic acid behaves similarly on bromination5.9 or iodination. lo Rather base RCH=CHCOzH + Br? +RCHBrCHBrCOzH --+analogous halodecarboxylations have been reported RCH=CHBr + COS + (HBr) (1) for p-aminobenzoic acid,11-13 anthranilic acid,14 pyrrole-a-carboxylic acids15 and e-furoic acids.I0 whose over-all result is the replacement of a carIt may be concluded from these references that boxyl group by halogen. It seemed of interest to study the mechanism of the more nearly direct re- halodecarboxylation occurs readily with aromatic carboxylic acids if the carboxyl group is located on placement of carboxyl by halogen a position of an aromatic ring which is readily susRCOzH Brz +RBr COn HBr (2) ceptible to attack by electrophilic reagents. I n We shall call this process i‘bromodecarboxylation.”3 other words the halodecarboxylations under conIn order to avoid the possible complication of sideration appear to involve electrophilic aromatic formation of intermediate dibromides (see eq. l), substitutions of a carboxyl group by halogen. Furaromatic rather than olefinic carboxylic acids thermore, halodecarboxylations probably involve have been studied. I n particular, 3,5-dibromo-2- reaction of the ionized form of the acid since such hydroxy- and 3,5-dibromo-4-hydroxybenzoicacids decarboxylations are retarded by strong acid and proceed readily with alkali metal salts of the acid.” were studied. Salicylic acid has been r e p ~ r t e d ~to- ~react with The present paper is concerned with the detailed bromine to form dibromosalicylic acid, which then mechanism of bromodecarboxylation as deduced may undergo replacement of carboxyl group by from a kinetic study of 3,5-dibromo-Z-hydroxy- and 3,5-dibromo-4-hydroxybenzoic acids in aqueous (1) Abstracted in part from the Ph.D. thesis of U. V. Henderson, acetic acid. Jr , Georgia Institute of Technology, April, 1954.
The stereochemistry and mechanism of the conversion of cis- and trans-cinnamic acid dibromides to /3-bromostyrene has been reported earlier from this Laboratory. These studies were concerned with the general process
+
+
+
(2) Papers I and 11, E. Grovenatein, Jr., and D. E. Lee, THIS 7 6 , 2639 (1953); E. Grovenstein, Jr., and S. P . Theophilou, ibid., 77, 3795 (1955). (3) This name is in accord with the suggestions for systematic nomenclature of substitution reactions as proposed by J . F. Bunnett and reported by A. M. Patterson, Chcm. Eng. News, 32, 4019 (1954). An alternative name, which has sometimes been used t o denote this reaction and also the Hunsdiecker silver salt reaction, is “brominative decarboxylation.” (4) M. A. Cahours, A n n . chim. phys., 131 13,8 7 (1845). (5) R . Benedikt, A n n . , 199, 127 (187U). ( 6 ) I. M. Kolthoff, Pharm. Weckblod, 69, 1159 (1932). (7) L. H. Farinholt, A. P.Stuart and D. Twiss, THISJ O U R N A L , 62, JOURNAL,
1237 (1940). ( 8 ) (a) E. J. Smith, Ber., 11, 1225 (1878); (b) E. Lellmann and R . Grothman, ibid., 11, 2724 (1884); (c) W. Robertson, J . Chem. SOL.,81, 1475 (1902); (d) R . B . Earle and H. L. Jackson, THIS J O U R N A L , 28, 104 (1906).
(9) F . G. Pope and A. S . Wood, J . Chem. SOL.,101, 1823 (1912). (IO) P. Weselsky, A n n . , 174, 99 (1874). (11) F. Beilstein and P. Geitner, i b i d . , 139, 1 (1866). (12) E. H. Wells, J . Assoc. Ofic.Agr. Chcmisls, 2 6 , 537 (1942). (13) H. L. Wheeler and L. M . Liddle, A m . Chem. J.,42, 441 (1909). (14) H . L. Wheeler and C. 0. J o h n s , i b i d . , 43, 398 (1910): F . UI1man and E. Kopetschni, Bcr., 44, 425 (1911). (15) Thus see A. H. Corwin, W. A. Bailey and I’. Viohl, THISJOURNAL, 64, 1267 (1942); H . Fischer, P . Halbig and B . Walach, A n n . . 462, 283 (1927). (16) Thus see H . B. Hill and G. T. Hartshorn, Ber., 18, 448 (1885); H. Gilman, H . E. Mallory and G. F . Wright, THISJOURNAL, S4, 733 (1932); A. P. Dunlop and F. N . Peters, “The Furans,’’ Reinhold PubIishing Corp., New York, N. Y., 1953, pp. 84, 116-121. (17) The reaction of bromine with a wide variety of sodium salts of carboxylic acids has been investigated recently from the standpoint of structural requirements by C. C. Price and J. D . Berman, A m . Chem. SOL.A b s t . , Cincinnati, March, 1955, p . 1 2 N.
570
ERLINGGROVENSTEIK, JR.,
AND
Experimental Detail@ Reagents.-Acetic acid (du Pont C.P. glacial grade) was purified by refluxing over chromium trioxide according t o the method of Bradfield and OrtonLgand then was distilled through a three foot, 13/8in. bore column packed with 3/10 in. glass helices. The fraction boiling a t 116-117" was collected and was analyzed for water content by titration with Karl Fischer reagent.m Acetic acid prepared by this method contained less than 1% water and the amount of water was known t o within 0.01%. The composition of acetic acid-water mixtures is always expressed in terms of percentage by weight. Six hundred milliliters of Baker analyzed C.P. bromine was refluxed for three hours over 60 g. of potassium bromide. The mixture was fractionally distilled and the middle cut boiling a t 57-53' was used in the kinetic runs.z1 2,4-Dibromophenol (Eastman Kodak Co., white label, m.p. 32.5-36') was recrystallized twice from chloroform a t -40" to give a product of m.p. 36.2-37" (recordedzz m.p. 40'). Z,&Dibromophenol (Eastman Kodak Co., white label) was distilled a t 18 mm. pressure and recrystallized from cold chloroform t o give a product of m.p. 5657' which agrees with that recorded.2z p-Hydroxybenzoic acid and salicylic acid were from Matheson and Eastman Kodak Co. and were recrystallized from water before use. Hydrobromic acid solutions were prepared from Matheson C.P. anhydrous hydrogen bromide; the concentration v a s determined by titration with standard barium hydroxide solution and this titration was checked in several instances by titration for bromide ion with standard silver nitrate solution with eosin as indicator. Lithium bromide solutions were prepared from Coleman and Bell C.P. lithium bromide and assayed by titration with standard silver nitrate solution. Baker and Adamson Reagent potassium dichromate, which was used as primary standard for assaying the sodium thiosulfate solutions, was recrystallized from hot water and dried i n vacua a t 156'. Other reagents were C.P. or Reagent Grade and were used without further purification. 3,5-Dibromo-Z-hydroxybenzoic Acid.-This cotnpouid mas prepared by bromination of salicylic acid in glacial acetic acid according to the procedure of Earle and Jacksonsd and was twice recrystallized from glacial acetic acid. Thirty grams of the product was refluxed for four hours in 1500 ml. of 1 N potassium hydroxide to saponify any esters which might have been formed in the bromination step. The mixture was neutralized with 200 ml. of 85y0 phosphoric iwid and filtered while hot. The precipitate was washed several times with water, sucked dry, and then shaken successively with 100-ml. and 50-mi. portions of n-pentane to extract any phenols present. The dried product was suspended in 300 ml. of water and titrated to a phenolphthalein end-point with ii ca. 0.1 S solution of barium hydroxide. The new solid formed in this process was indicated hl- the titer of base t o he the barium salt formed b y ueutralization of one acidic hydrogen per molecule of dibromosalicylic wid. The barium salt was filtered, washed with water, then decomposed by triturating xith 120 ml. of 2 N hydrochloric acid. The dibromosalicylic acid so obtained was filtered, washed three times with water, dried in air, recrystaliized from acetouitrile and finally dried in FUCZLO a t 56' to giw n product of m.p. 228-230' (m.p. recordedz3228"j. 3,5-Dibromo-4-hydroxybenzoic Acid.-This compound was prepared by bromination of p-hydroxybenzoic acid ;ICcording t o the procedure used above for 3,5-dibromo-2hydroxybenzoic acid. The product after one recrystallization from glacial acetic acid was obtained in 55% yield arid had m.p. 275-276' (m.p. recorded2' 267-268"). The product was subjected to about the same purification procedure (18) All melting points and boiling points are uncorrected unless otherwise specified. (19) K. J. P. Orton and A. E . Bradfield, J . Chem. SOL.,960 (1924). (20) J. illitchell, Jr., and D. M. Smith, "Aquametry," Interscience Publishers, Inc., New York, X. Y., 1948. JOURNAL, 73, (21) Cf. H. G. Kuivila and E. K. Esterbrook, TIXIS 4629 (1951). (22) I. 31. Heilbron and H. 3T. Bunbury, "Dictionary of Organic Compounds," Oxford 7Tniversity Presq, Xew York, N. Y . , 19.43. Vc,l. 1 , p . 713. (2:3) Reference 22, p . 7 2 0 . 124) Referenre 22, p . i f l o .
U. V.HENDERSJX, JR.
VOl. 7 8
as used above for 3,5-dihromo-2-hydroxybenzoicacid cxcept that the barium salt formed from neutralization of both acidic hydrogen atoms of the $-hydroxy acid was water soluble and was precipitated by addition of ethanol. A final addition crystallization from chloroform gave a product which still had m.p. 275-276'. AnaZ.25 Calcd. for C7H403Br2: Br, 54.0. Found: Br, 54.3. Kinetic Measurements.-A stock solution of solvent of the desired water-acetic acid composition (reported on the basis of the percentage by weight of acetic acid present) was prepared so as to contain the required concentration of electrolytes. IVith this solvent 250 ml. of solution of the compound being brominated was prepared in the desired concentration (0.0025-0.0125 M ) and either 20.0 or 40.0 ml. of the solution was placed in each of a number of 50-ml. or 100-ml. red, low actinic, volumetric flasks fitted with ground glass stoppers. A like volume of the solvent was also placed in each of several volumetric flasks t o serve as blanks. A bromine solution was prepared by adding a few tenths of a milliliter of bromine t o 100 mi. of the solvent. All solutions were kept a t the temperature of the waterbath which mas 20.0 & 0.1". The reaction was started by pipetting a t a noted time 5.0 or 10.0 ml. of the bromine solution into each of the prepared flasks so as t o make, respectively, 25.0 or 50.0 ml. of reaction mixture. The reaction was stopped a t the desired time by injecting with a syringe 4 ml. of 1 potassium iodide solution followed b y 10 ml. of 4 N potassium hydroxide ( t o decrease the acidity of the solution and help prevent air oxidation of iodide ion). The solution was rinsed into 10 ml. of water and the iodine equivalent of the unreacted bromine was determined by titration with standard sodium thiosulfate solution t o Visual disappearance of the iodine color. I t seemed possible that the iodine (or triiodide ion) formed in stopping the reaction might itself continue t o react with the remaining dibromohydroxybenzoic acid in the solution. This was proved to he unimportant since titers upon runs with dibromo-P-hYtlroxybenzoic acid, in vihich the reaction mixture after addition of potassium iodide and sodium hydroxide was allowed to stand for 20 to 30 min. before titration, were not apprcciably different from those which were run similarly but were titrated immediately (within 1t o 2 inin.) after stopping the reaction. Apparent second-order rate constants (kspp) were calculated from each point by use of the integrated rate equation
k,,,
2 303 lop n(b - x ) b(a - x )
= -__ ( b - a)t
irhere k,,, is expressed in units of liter mole-' secorid-', is the initial molar concentration of substrate bcing brominated, b is the iuitial molar Concentration of bromine (stoichiometric concentration as determined by thiosulfate titration), and x is the molar concentration of substance reacted (or product formed) a t time t . Bromination Products.-In preliniinary experiments upon bromination of dibromosalicylic acid in 40% by volume dioxane in a chloroacetic acid buffer a t pH 3.4, "tribromophenol bromide"6**6was obtained as a light yellow precipitate. Tribromophenol upon bromination under similar conditions gave the same product. I n the same sol\-elit hut 0.1 N in perchloric acid, tribromophenol bromide formation was greatly retarded and 50y0yield of tribromophenol was obtained (by isolation) upon bromination of dibromosalicylic acid with excess of bromine. S o evidence for tribromophenol bromide formation was obtained during the present brominations conducted in aqueous acetic acid as solvent with hydrobromic or perchloric acid concentrations as low as 0.05 M . Infinity titers upon bromine consumption indicate that dibromosalicylic acid and dibromo-p-hydroxybenzoic acid react with exactly one molar equivalent of bromine under the conditions of the present kinetics. Thus dibromosalicylic acid (in 70.0% HOAc. 0.300 X in H B r ) gave 99.570 of this theoretical amount of
Feb. 5, 1956
BROMODECARBOXYLATION OF BROMOHYDROXYBENZOIC ACIDS
bromine consumption in 256 hours and nine runs upon dibromo-p-hydroxybenzoic acid (in 80% acetic acid, 0.1 or 0.3 M in HBr) gave infinity titers which averaged 99.8 f 0.5% of this theoretical consumption of bromine. The amount of carbon dioxide evolved daring the bromination of dibromo-p-hydroxybenzoic acid ixa 80.2% acetic acid, which was 0.10 M in both perchloric acid and lithium bromide, was determined by measuring the volume of carbon dioxide evolved from the well shaken solution which had been saturated with carbon dioxide before breaking a capsule of bromine in the solution. Two such experiments indicated that about 97% of the theoretical amount of carbon dioxide was evolved. Finally tribromophenol was isolated in 23% yield from dibromosalicylic acid under similar conditions of concentration as in the kinetic runs (75% acetic acid, 0.100 M HBr, 0.200 M LiBr) but a t ca. 40" and similarIy in 55% yield from dibromo-p-hydroxybenzoic acid (80% acetic acid, 0.300 dl4 HBr). Equilibrium Constant for Tribromide Ion Formation.The equilibrium constant K1 for the reaction Brz Br- = Br3-, as defined by
+
K I = [Bra-]/[Brzl [Br-]
(3)
where K , is expressed in liters/mole, was determined by the general method of Popov, et eZ.,27with use of a model DU Beckman quartz spectrophotometer thermostoted a t 20 '. The solutions were contained in ground-glms stoppered Corex cells of 1.000 f 0.001 cm. optical path length. All solutions were made up to an ionic strength of 0.300 by addition of appropriate amounts of perchloric and hydrobromic acids. The bromine concentration of the solutions was determined by titration as described in the kinetic measurements. The molar absorbancy indices28 in the solvents 70.0, 75.0 and s O . O ~ oacetic aci& were within experimental error not measurably different over this range of solvent composition and the average values of these indices are recorded in Table I.
TABLE I MOLARABSORBAX'CYINDICES FOR BROMINEAXD BROMIDE IONIN 70-80% ACETICAUD AT 20" Bromine
ax
360 380
106 165
400
179
420
159
571
Results Preliminary experiments of Table 11-runs 1 and 2 for dibromosalicylic acid; runs 8, 9, 13 and 14 for &bromo-p-hydroxybenzoic acid-indicated that for comparisons a t equal hydrogen and bromide concentrations the bromodecarboxylations were first order in both the stoichiometric concentration of bromine and that of the dibromocarboxylic acid. TABLE I1 KINETIC DATA FOR Sulvent, wt.
?6
103. (Ar-
108.
BROMODECARBOXYLATION AT 20.0" (HCI-
Hdo, (flrda, Odo,
HO- moles/ moles/ mole/ Run Ac 1. 1. 1.
iHBr)o, (LiBrh, mole/ mole/ I. 1.
102kapp,
]./mole sec.
ArH2 = 3,5-Dibromo-2-hydroxybenzoic acid 1 2 3 4 5
75.0 75.0 75.0 75.0 75.2
5.01 5.01 5.01 5.08 5.03
3 . 7 6 0.100 7.50 ,100 7.44 ,100 4.86 ,100 3.93 ,100
0.0O4Zc .0040d ,102
5.8' 5.8 4.7 5.0 0.49
ArHz = 3,5-Dibromo-4-hydroxybenzoicacid 6 7
75.0 80.0 80.2 80.2 80.2 80.2
8 9 10 11 12 13 14 15
80.0 80.0 80.0 80.0
5.00 5.01 4.97 5.05 6.69 4.61 4.91 5.18 8.93 4.46
5 . 7 9 0.100 5.82 ,100 4.17 ,100 0.103 8.12 ,100 ,102 5.57 ,100 .095 ,104 7.87 ,050 8.02 0.107 8.05 ,222 5.10 ,2225 8.14 ,200 .lOOC
200 63 1.29 1.31 1.52 3.4 0 . 7 2 =t0.05b ,064 zt .OOlb .OOlb ,061 ,060 .001*
+
+
The values of kapp are initial rate constants (obtained by extrapolation) unless otherwise specified. * Average value of Rap, and mean deviation. c KBr rather than LiBr. KCI'OI rather than LiBr. TRI-
Tribromide ion QT
966 755 556 309
The absorbancy index of bromine in these 0.300 M perchloric acid solutions is essenikdly the same as that reported" a t 400 mp in glacial acetic acid. Furthermore the absorptivity of bromine in 80% acetic acid, 0.300 A4 in lithium bromide, is indistinguishable from that in the same solvent 0.300 M ; in hydrobromic acid. These results show that the bromine spectrum, while dependent upon bromide ion concentration, is relatively indepenaent of ehe concentration of strong acids. This fact argues against the presence of appreciable amounts of any bromine species whose concentration is acid dependent. From the data a t each of t h e four wave lengths shown in Table I and from the absorbance of solutions 0.003-0.005 M in bromine, 0.0010 to 0.0015 M in hydrogen bromide, and containing enougH perchloric acid to make the solutions 0.300 M in strong acid, the equiiibrium constant (expressed in terms of concentration) for tribromide ion formation was calculated to be 84 in 70.0%, 88 in 75.0%, and 92 l./mole in 80.0% acetic acid, with an average deviation of the individual values from these mean values of +a%. The value obtained in!757a acetic acid may be compared with the value of 91 =k 3 reported30 in the same solvent a t about 0.014.02 ionic strangth by an aspitation technique. (27) A. I . Popov, K. C. Brinker, L. Campanaro and R. W. Rinehart, THIS JOURNAL, 7S, 514 (1951). (28) For spectrometric nomenclature and detlnition of terms see National Bureau of Standards, Letter-Circular LC-857 (1947). (29) R. E. Buckles and J. F . Mills, T H i 6 JOURNAL, 78, 552 (1953). (30) A . E. Bradfield, G. I . Davies and E. Long, J . Chem. Soc., 1389 ( I 0-L!)) .
a
Hence apparent second-order rate constants (kapp) were calculated on this basis. These constants are highly dependent upon the hydrogen and bromide ion ccmcentfations as is shown by the exploratory runs of Table I1 and, by the more detailed experiments of Table 111, which latter were all performed a t ionic strength of 0.300. In view oi this depend14.0 [ -
6 .O
h
20 40 60 Reaction, %. Fig. 1.-Integrated second-order rate constants 8s. percentage reaction for 3,5-dibromo-2-hydroxybenzoicacid (0) in 75.0% acetic acid, (HBr)o = 0.0100 J!!, (HClOa) = 0.290 iW; 3,5-dibromo-4-hydroxybenzoic acid (0)in SO.O'% acetic acid, (HBr)" = 0.050 M , (HClO,) = 0.250 M .
0
572
ERLINGGKOVENSTEIN, JR.,
AND
VOl. 78
U. V. HENDERSON, JR.
TABLE I11 KINETICDATAFOR BROMODECARBOXYLATION AT 20.0", (ArH2)" !2 0.005 M , p = 0.300 lO'(Brdo, moles/l.
(HClOdo, mole/l.
(HBr)o, mole/l.
(LiBrh,
101 [Bra-] 8 , moles/l.
mole/l.
102k*, lO'k', (moles/l.)z set.-' I./mole sec.
lO'kapp,
l./mole sec.
3,5-Dibromo-2-hydroxybenzoic acid in 75.0Yc acetic acid 7.94 7.93 8.00 7.97 7.95 8.37 7.96 7.90 8.12 7.87
0.300 .290 .250 .200
.loo
0.0100 .0500
.loo
.200 .300 .200
0.100 .200
.loo
25.5" 13.5" 3.47" 1.37" 0.440 i.0 ,002b .233 =k ,008 ,428 =k ,003 1 . 1 5 i .02 1.18zk .02c 11.0293 i .0003d
3.02 6.34 7.10 7.50 8.05 7.67 7.60
.300 .300
2.55 2.18 1.68 1.25 0.79 0.62 1.14 3.07
0.137 0.66 1.05 1.37 1.64 1.34
0.90
3,5-Dibro1~io-4-hydroxybenzoicacid i n 80.0% acctic acid 8.19 0.300 (0.42)O 300" ( 0 . 16)e 30 6.50 7.6" 1.49 3.80 8.13 .250 0.050 8.17 ,200 ,100 7.31 1 . 8 2 i 0.03b 1.45 1.73 8.24 ,100 .200 7.80 0.50 zk .01 1.60 0.92 1.62 0.62 8.06 .300 7.78 ,222 & ,003 8.25 .200 0.100 7.96 .52 f .01 1.70 1.45 8.12 .loo ,200 7.82 2.00 f .04 1.63 5.58 8.16 .050 ,250 7.86 6 . 9 zk .2 1.40 19.2 8.04 ,300 1 . 9 7 f .04' .lC 7.51 ,300 11.6 z t Initial rate constant (extrapolated). Average value of k,,, and mean deviation. In 70.0% acetic acid. In 80.0% acetic acid. * Calculated on the basis of an assumed stoichiometric concentration of Br- of 0.0010 M. f I n 75.0% acetic acid.
ence, the individual apparent second-order rate constants calculated for each run generally decreased with time since hydrogen bromide is formed as a product of the bromodecarboxylations as shown by equation 2. This decrease was, to be sure, largest for those reactions with small initial hydrogen or bromide ion concentrations and was almost negligible when both of these concentrations were 0.3 M . When the fall in kapp was 8% or less (from 0 to 70% reaction) during a run, the individual values of kapp were averaged (thus see Table IV); in other cases the values of kapp were obtained by extrapolation of the individual values of k to zero percentage reaction (thus see Fig. 1). TABLE IV BROMODECARBOXYLATION O F 3,5-DIBROMO-IHYDROXYBENZOIC ACIDIN 80% ACETICACIDAT 20.0" (ArH2)o = 0.005128 M , ( R r Z ) ~= 0.008240 Af, (HBr)o = 0.200 M , (HCIOd)o = 0.100 M ; 50.0 ml. of reaction mixture titrated with 0.02185 df Sa&Os KINETICS
OF
Time X 10-3, sec.
Titer, mi,
0 25,32 76.14 175.3 263.9 360.8 446.6
37.71 35.41 31.92 27.28 24.36 21.82 21.00
X 108, l./mole sec.
kapp
0.511 ,492 ,484 ,492 .520 ,474
Reaction.
%
0 9.8 24.7 44.4 56.9 67.6 71.2
Av. ,496 i 0.014
For the sake of comparison, the kinetics of the bromination of 2,6-dibromophenol was studied as recorded in Table V, in which also are recorded a few runs upon 2,4-dibromophenol.
TABLE
KISETICDATAFOR BROMODEPROTOSATION ~r 30 0' 108. (Brh moles) 1.
(HCI04)o
mole'/ 1.
loa. ( H B r h , (LiBr)o, [Br3-lo. mole/ 1.
mole/ 1.
moles/
knpp,
1.
I./mole sec.
k', I./mole sec.
2,6-Dibromophenol, (Ar0H)o E 0.0025 AT, in 80.0% Acetic Acid p = 0.300 3.71 3.75 4.06 3.73 7.98 3.69
0.200 ,100
0,100 ,200 ,300 ,200 ,150 ,100
0.100 ,150 ,200
3.31 3.56 3.92 3.60 7.70 3.56
0.513 i0.004* , 2 4 4 + ,001 ,160zk , 0 0 3 . 1 9 9 L ,003 , 2 3 2 & ,003 ,302 & , 0 0 8
5.08 4.66 4.52 5.61 6.48 8.53
2,4-Dibromophenol, (Ar0H)o !2 0.005 15f 3.77 7.47
0.100 ,100
0.100 ,137
0 . 53'gC 24b,d
Average value of k,,, and iiiciiti deviation. * Initial rate constant (extrapolated). hi 72.2y0 acetic acid. In 80.2% acetic acid.
Discussion 3,5-Dibromo-4-hydroxybenzoic Acid.-This
acid underwent bromodecarboxylation a t a rate such that kapp was approximately inversely proportional to both the square of the bromide ion and the square of the hydi-ogen ion con~entrations~l over the range of about 0.05 to 0.3 M in each ion. This is illustrated by the near constancy of R' of Table 111, where k' = k,,,, K~[Br-l*(Brd [H +l2/[Bra-1
(4)
III this equation as elsewhere in this paper the rntities enclosed in brackets refer to the actual coil-
centration of the species shown, while those in pa(31) It is assumed throughout the present work that the hydrogen ion concentration is equal t o the sum of the molar concentrations of HBr and HCIO, and thus that any contribution of the acetic acid t o the hydrogen inn concentration is negligible lor these soltitions which were ulways 51 lPart 0 05 $1 in mineral acid.
Feb. 5 , 1956
BROMODECARBOXYLATION OF BROMOHYDROXYBENZOIC ACIDS
rentheses refer to the stoichiometric concentration of a material, in this case the concentration of bromine as determined by thiosulfate titration. I n view of the magnitude of the bromine-tribromide ion equilibrium constant, K1, as defined by equation 3, 80% or more of the bromine is present as tribromide ion under the present conditions. The term K1[Br-](Brz)/[Br3-] is accordingly used in equation 4 to correct kapp for bromine which has been converted to unreactive32 tribromide ion, in other words rate = k’[Brz](ArHz)/[Br-] [H+I2
( 5)
where (ArH2) is the stoichiometric concentration of dibromo-p-hydroxybenzoic acid. The inverse squared hydrogen ion dependence and the inverse bromide ion dependence imply that the transition state complex of the rate-determining step has two less protons and one less bromide ion under the reaction conditions specified than do the assumed reactants. One possible mechanism fulfilling these requirements is (mechanism I)
+ HzO =HZOBr+ + BrKz
Brz
[HzOBr’]
= Kz[Brp]/[Br-]
Ka ArHz
Ar--
+ 2H+
[Ar--] = Ka[ArHx]/[H+]2 rate = k”[H20Br+][AI---]
(6)
(6‘) (7) (7’) (8)
Then from substitution of (6’) and (7’) into (8) rate = k“K~Ka[Brz] [ArHz]/[Br-] [H+IZ= k’[Br~l[ArHzl/[Br-I [ H + l z (9)
Since in a medium as acidic as the present the concentration of un-ionized dibromo-p-hydroxybenzoic acid must be equal essentially to the stoichiometric concentration of this acid, then mechanism I leads to a kinetic result (equation 9) identical with that found (equation 5 ) . It should be noted that somewhat different brominating agents, Br+ and CHaCOzHBr+ (from acetolysis rather than hydrolysis of bromine) or any combination of these and hypobromous acidium ion (H20Br+), lead also to the observed kinetic result. Finally the present kinetics cannot distinguish between protons having been lost from the brominating agent or from the dibromo-p-hydroxybenzoic acid. Thus the observed kinetics are also in agreement with a ratedetermining interaction between hypobromous acid and mono-ionized dibromo-p-hydroxybenzoic acid or between hypobromite ion and un-ionized dibromo-p-hydroxybenzoic acid. The former of these possibilities seems improbable under our conditions of acidity on the basis of studies upon the bromination of substances which cannot undergo complicating ionizations. 32,33 These studies indicate that hypobromous acidium ion (or brominium ion, Br+) is a much more reactive brominating agent than hypobromous acid and on general evi(32) Thus see D. H. Derbyshire and W. A. Waters, J . Chcm. Soc., 564 (1950). (33) W . J. Wilson and F. G. SoDer. ibid.. 3376 (1949): . ,. E.Shilov -~ - and N. Kanyaev, Com9t. rend. acod..sci, U.R.S.S., 2 4 , 890 (1939); Chcm. Abst., 34. 4062( 1940); S.J. Branch and B . Jones, J . Chcm. Soc., 2317 (1954).
573
dence hypobromite ion should be even less reactive than hypobromous While the kinetics of the bromodecarboxylation of dibromo-p-hydroxybenzoic acid fit equation 4 or 5 moderately well in the range of bromide and hydrogen ion concentrations of 0.04 to 0.3 M , the computed values of k’ show some tendency to diminish (by as much as some 15%) toward the lower limits of these concentrations. Moreover when no bromide ion was initially added to the reaction mixture the estimated value of k’ was no more than about a tenth of the values previously calculated. This estimation involves an assumption as to what the effective concentration of bromide ion was corresponding to the value of kapp, here obtained by extrapolation back to zero per cent. reaction. T o the extent that this extrapolation is accurate, the correct value to use for the bromide ion concentration is to be computed from the solvolysis constant of bromine. From the hydrolysis constant of bromine136the bromide ion concentration may be computed to be 10+ M for water as solvent; a lower value doubtless is attained in 80% acetic acid. As a much more conservative estimate a value of lo-* M was used here for the stoichiometric bromide ion concentration in the calculation of k’ (Table 111) ; this bromide ion concentration corresponds to the bromide formed by 25% reaction and thus should lead to an estimate of a maximum value for k’. That this value of k’ is considerably smaller than the values calculated a t higher bromide concentrations leads to the conclusion that mechanism I , in any of its modifications mentioned above, is basically incorrect. This view is substantiated by the fact that in previous brominations in water or in aqueous acetic acid in which molecular bromine was used as the brominating agent, the kinetic evidence has been that the effective brominating agent was Brz,33,a6 sometimes accompanied by ( B I - ~ ) ~ .These ~’ previous brominations, i t is true, involved electrophilic substitution of bromine for hydrogen rather than carboxyl, but i t is difficult to see why the relative order of activity of brominating agents should vary greatly with the nature of the displaced group (barring steric effects) provided that aromatic halogenation follows the general pattern of nitrationls8namely, that the entering group has become more or less fully bonded before the bonding of the old group is substantially broken. Further discussion of the mechanism of bromodecarboxylation of dibromo-p-hydroxybenzoic acid is deferred until after a discussion of 2,6-dibromophenol and dibromosalicylic acid. 2,6-Dibromophenol.-In order to compare more fully the mechanism of bromodecarboxylation with that of bromodeprotonation, the kinetics of the bromination of 2,6-dibromophenol has been studied (34) For a general survey of aromatic electrophilic halogenation see C. R. Ingold, “Structure and Mechanism in Organic Chemistry.” Come11 Univ. Press, Ithacs, N. Y.,1953, pp. 288-295. (35) H. A. Liebhafsky, THISJOURNAL, 66, 1500 (1934). (36) A. E. Bradfield, B. Jones and K. J. P. Orton, J . Chcm. SOC.. 2810 (1929). (37) A. E. Bradfield, G.I. Davies and E. Long, {bid., 1389 (1949); P.W. Robertson, P. B. D . de la Mare and W. T. G. Johnston, ibid., 276 (1943). [38) See ref. 34, pp. 279-281.
674
ERLING GROVENSTEIN, JR.,
AND
U. V. HENDERSON, JR.
1701.
78
under the same general conditions as those used for 3,5-dibromo-4-hydroxybenzoic acid. The available data are given in Table V. The apparent second-order rate constant (kapp) can be corrected for bromine which has been converted t o tribromide ion by the relation
eliminate the possibility that the bromodecarboxylation of dibromo-p-hydroxybenzoic acid proceeds by way of a fast, presumably acid-catalyzed, protodecarboxylation to give 2,6-dibromophenol followed by a rate-determining bromination of the latter. Such a mechanism would demand that the rate of the apparent bromodecarboxylation be the same as ksppK1(Brp)[Br-]/[Brs-] k* (10) the rate of bromination of 2,6-dibromophenol. At constant hydrogen ion coneentration, the values The possibility of a slow rate-determining protoof k* are constant within 11% over a threefold decarboxylation of dibromo-p-hydroxybenzoic acid variation of bromide ion concentration as can be followed by a fast bromination of the resulting seen from the table. This bromide ion dependence phenol is, of course, eliminated by the kinetic deindicates that molecular bromine (Bra) is the effec- pendence upon bromine concentration shown by the tive brominating agent. Such variation as exists acid. Similar considerations hold for the bromoin k* here may be the result of specific salt &ects; decarboxylation of 3,5-dibromo-2-hydroxybenzoic an alternative interpretation will be given later. acid which is brominated a t an apparent secondA plot (Fig. 2) of K* at constant bromide ion order rate of less than one-hundredth that of 2,4concentration versus the reciprocal of the hydrogen dibromophenol as is shown by the data of Tables I1 ion concentration reveals a linear relationship with and V. k* increasing as the hydrogen ion Concentration 3,S-Dibromo-2-hydroxybenzoicAcid.-This acid decreases. The data therefore fairly closely fit the underwent bromodecarboxylation a t a rate such relationship that neither k' nor k* (as previously defined) were K,,,K1(Br?)[Br-]/[Brs-] = K O K-/fH+] = k * ( l l a ) constant a t constant hydrogen ion concentration (see Table 111). Moreover the trend of these valor ues indicates that kapp follows a non-integral brorate = ka,,(ArOH)(Brz) = (ko(Ar0H) mide ion dependence, somewhere between an ink- (ArOH) / [H+I ) [Brs-)/KI [Br -1 (1 lb) verse first- and second-power dependence. In where (ArOH) is the concentration of the 2,6-dibro- other words bromination of dibromosalicylic acid mophenol, ko equals 2.5 l./mole sec., and k- equals (3,5-dibromo-2-hydroxybenzoicacid) follows a bro0.60 set.-'. Since the term [ B r ~ - l / K ~ [ B r -equals l mide dependence which lies in between that previthe concentration of bromine, equation ( I l b ) corre- ously generally noted for 3,5-dibromo-4-hydroxysponds to the expected kinetic equation for the benzoic acid and for 2,6-dibromophenol. This attack of molecular bromine upon ionized and un- result suggests that perhaps dibromosalicylic acid ionized 2,B-dibromophenol (mechanism 11). reacts by a mixture of mechanisms I and 11. Such a possibility leads t o the rate expression
+
+
rate = k"[HzOBr+][Ar--]
+ ko[Brs][ArHz] +
k- [Brs][ArH-1
r
z 4.0 T i
2.0
0.0 2.0
4.0 6.0 8.0 k*, l./mole sec. Fig. 2.--Variation of k* with l / [ H + ] for 2,g-dibromophenol a t [Br-] equal 0.292-0.296 M (0),0.196 M ( 8 ) . and 0 097 M (6). Points drawn with radius equal 2% of k*.
2,6-Dibromophenol undergoes bromination in 80% acetic acid to give an apparent second-order rate constant which is some 150 to over 700 times greater than that of dibromo-p-hydroxybenzoic acid under identical conditions of hydrogen and bromide ion concentrations as can be seen by comparison of the values of kapp in Tables 111 and V. The rate of bromination of the phenol relative t o the acid increases as the concentration of hydrogen and bromide ions increases. These data clearly
where [ArH2], [ArH-] and [Ar--1 represent the concentration of dibromosalicylic acid in its unionized, mono-ionized and di-ionized forms, respectively. Terms for the attack of H20Br+ (as well as any species of similar bromide ion dependency as discussed under mechanism I) and Brz upon all three of these forms of dibromosalicylic acid might be included in the above rate expression, but a t constant hydrogen ion concentration, in any event, all such expressions reduce to rate = kA[ArHz][Brz]
+ krrr[ArHz][Brp]/[Br-]
(12)
where k"' and k ; include any hydrogen ion dependency. An analogous mixture of mechanisms has been proposed to account for the iodination of tyrosineraehistidineI40 as well as certain pyrroles.41 As a test of this mechanism for dibromosalicylic acid, since by definition of kapp rate = k,,,(ArHJ(BrJ
(13)
and since under our conditions of acidity, to a good approximation (ArHz) = [ArH*], then kapp(BrJ/[BrJ = k* = k;
+ A"'/[Br-l
(14)
Hence for this mixed mechanism to hold, a plot of k* vx. l/[Br-] should give a straight line. Such a plot a t constant hydrogen ion concentration is (39) C.H. Li, THISJOURNAL, 64,1147 (1942);66, 228 (1944). (40) C.H.Li, i b i d . . 66, 225 (1944). (41) K.W. Doak and A. H. Corwin, i b i d . , 71, 159 (19.49).
Feb. 5 , 1956
575
BROMODECARROXYLATION OF BROMOHYDROXYBENZOIC ACIDS
given in Fig. 3 and the marked deviation from linearity argues strongly against this mechanism for the bromodecaxboxylation.
(21)
Since (22)
rate = ks[ArBr-] = ka,,(Br2)(ArHz)
substitution of (21) into (22) gives
120
ir Application of equilibrium statistics and the principle of microscopic r e ~ e r s i b i l i t yto~ reactions ~ 15, 16, and 17 leads to the relationship (24)
kiK4Ik-i = kZ/k-g
which must hold even away from equilibrium. From (23) and (24) there may be obtained l/K*[H+]* = k - ~ [ B r - ] / k ~ K * k ~ KI/{ ~ k*[H+l2
+
+
1.0 2.0 3.0 look*, l./niole sec. Fig. 3.-Variation of k * with l/[Br-] for 3,5-dibromo2-hydroxybenzoic acid a t [H+] = 0.300 AT; points drawn with radius equal 2 % of k*.
0
The following mechanism (111)is therefore postulated to account for the observed bromide ion and hydrogen ion dependency of k.,,,
According to (25) a plot of 1/k*[H+l2 os. [Br-] should give a series of parallel straight lines for each hydrogen ion concentration. Such a plot for dibromosalicylic acid is given in Fig. 4 in which the radii of the points are drawn equal to 3% of I l k * [H+]*save for the point at zero bromide ion concentration where the radius is 5y0. Extensive data are available only a t 0.300 M hydrogen ion concentration; the data here do indeed fall rather well upon a straight line all the way down to an initial bromide ion concentration of essentially zero.
fast
ArHz
zzArH- + H + ; Ka = [H+][ArH-]/[ArHJ (15) ArH-
ki
+ Brz
ArHBr
kkz
ArH:
+ Br-
0'30
r--lF-T /
/
/
/
/
/
I
+ Br? I_ ArHBr + H + + Brk-
(16)
kiK~[H+l I (25)
(17)42
z
fast
ArHBr
t ArBr- + H + ; Kj = [ArBr-][H+]/[ArHBr] ( 18) ArBr-
k3
+ Ar'Br-
+ COZ
( 19)
where Xr'Br- is the anion of tribromophenol. Here the reactive intermediate ArHBr is presumed for dibromosalicylic acid to have the structure I.
I""
BrVC02H
( 1)
Br / B r
This intermediate ArHBr is assumed under the conditions of the kinetics to be formed in low concentration relative to ArH2. Then on the basis of the steady-state approximation
10 20 30 1/100k*[H*I2, sec. l./mole. Fig. 4.-[Br-] os. 1/100k*[H']2 for 3,5-dibromo2-hydroxybenzoic acid a t [H+] equal 0.300 AT (0),0.200 M (a), and 0.100 M ( 8 ) . 0
At negligible bromide ion concentrations equation 25 reduces to k * = kz
+ klK,/[H+]
(26)
Substitution for [ArH-] from (15) and for [ArHEr] from (18) and solving for [ArBr-] gives
which equation is entirely analogous to that ( l l a ) obtained with 2,6-dibromophenol. By extrapolation of the data a t 0.200 and 0.100 M hydrogen ion concentrations to zero bromide ion concentration as suggested by equation 25 and as shown in Fig. 4, values of k* a t zero bromide ion concentration
(42) T o be sure the proton formed here is affixed to some base; this itep ia presumed to be evbject t o general base catalyhis.
(43) Thus see A. A . Frost and R .G. Pearson, "Kinetics and Mechanism," John Wiley wnd Soas, Inc., Now York, N . Y . , 1053,p . 202,
+
+
kl[ArH-] [Brs] k?[ArHz][ B ~ z= ] k-~[ArHBrj[Br-] k-z[ArHBr] [Hi] [Br-] ka[ArBr-] (20)
+
576
ERLINGGROVENSTEIN, JR.,
AND
may be obtained and these upon substitution into (26) permit approximate evaluation of k1K4 and k2, which are found i o be, respectively, 0.0042 rt 0.002 set.-' and 0.0103 f. 0.0005 l./mole sec. From this value of k1K4 and the slope of the lines of Fig. 4, k--]/k3K6 is calculated to be 20 (l./mole)2 and k-Z/k& 49 ( l . , ' m ~ i e ) ~ . Rediscussion of 3,5-Dibromo-4-hydroxybenzoic Acid.-The above discussion upon dibromosalicylic acid suggests that perhaps mechanism I11 also applies to dibromo-p-hydroxybenzoic acid. A plot of l/k*[H+I2 us. [Br-] a t a hydrogen ion concentration of 0.300 M gives a good straight line for this acid. It is notable that the point a t zero initially added bromide ion falls on this line. This latter point is much nearer the origin than the corresponding point for dibromosalicylic acid; this means that the second term on the right of equation 25 is quite small compared t o the first term. To the extent that the second term is negligibly small, then equation 25 reduces to equation 4 which was previously shown to hold approximately for dibromo-p-hydroxybenzoic acid. Mechanism 111, therefore, offers an alternative and superior explanation for this kinetic form to that considered under mechanism I.44 The points for smaller hydrogen ion concentration for dibromo-p-hydroxybenzoic acid fall much closer to the line for 0.300 M hydrogen ion concentration than did the corresponding points for dibromosalicylic acid. This makes evaluation of k& and kz difficult for the p-hydroxy acid. If the point a t 0.05 M hydrogen ion concentration is considered to be accurate, then k1K4 and kz are found by the method previously discussed to be approximately 0.05 set.-' and 0.12 l./mole sec., respectively; from these values and the slope of the line, k-l/k& and k-Z/k3K5 are, respectively, -300 (l./mole)2 and -700 ( l . / m ~ l e ) ~ On . the basis of these constants, errors as small as 6% on the evaluation of kapp can account for the rather inconsistent location of the points a t 0.2 and 0.1 M hydrogen ion concentration. A comparison between the kinetics of bromodecarboxylation of 3,5-dibromo-2-hydroxy- and 3,5dibromo-4-hydroxybenzoic acids is of interest. An obvious point of distinction is that the apparent second-order rate constant for the p-acid is 7.6, 8.5 or 9.8 times that of the o-acid depending upon whether the comparison is being made in 80, 75 or 7oy0acetic acid a t a hydrogen bromide concentration of 0.300 M . According to equations 25 and 1l a the apparent second-order rate constants at a given bromine, bromide and hydrogen ion concentration are a moderately complex function of the constants k l , k - ] , kS, k3, k4 and Kd. While these constants have not yet all been evaluated separately and while those for the p-acid refer to 80% and those for the o-acid to 75y0 acetic acid, certain correlations are possible. I n these two solvents a t hydrogen bromide concentration of 0.3 M the oevidence f o r mechanism I11 for 3.5-dibromo-4hydroxyhenzoic acid recently has been provided by study of carbon-13 isotope fractionation a* a function of bromide ion concentration for the bromodecarboxylation (Gus A. Ropp and E. Grovenstein, Jr.. work done at t h e Oak Rirlce h-ational 1 ahoratory; details t o be reported later).
U. V. HENDERSON, JR.
voi. 78
and p-hydroxy acids undergo bromodecarboxylation a t nearly the same rate. This is due to the fact that while k,KI and kz for the p-hydroxy acid are some ten times larger than the corresponding constants for the ortho-hydroxy acid, this effect is approximately compensated by some 15-fold larger values for the reversibility ratios, k--l/k&6 and k-Z/ksK6 for the p-hydroxy acid. Upon going from 80 to 75% acetic acid k& and k p would both be expected to increase for the p-hydroxy acid (the former largely because of increase in K 4 ) and k-l/ k3K5 and k--2/k3K5 would be expected to decrease (largely because of increase in K6 and decrease in k-2) on the basis of current theories46of solvent effects. I n this way, therefore, the relative reactivities of the 0- and p-hydroxydibromobenzoic acids in 75y0 acetic acid may be rationalized in terms of mechanism I11 ; similar explanations apply also to the relative reactivities in the other solvent mixtures. A point worth emphasizing is how large these solvent effects are. Thus in presence of 0.3 M HBr for dibromo-p-hydroxybenzoic acid, kapp increases 8.9-fold in going from 80 to 75% acetic acid and 5.9-fold from 75 to 70% acetic acid; the corresponding factors for dibromosalicylic acid are 8.0- and 5.1-fold. Such large solvent effects demanded very accurate preparation of the mixed solvents and doubtless were a major source of error in our present work. Salt effects are expected to be correspondingly large for the present reactions and for this reason quantitative comparisons have all been made a t a large and constant ionic strength of 0.300. That k1K4 and kz in a given medium should be larger for dibromo-p-hydroxybenzoic acid than for dibromo-o-hydroxybenzoic acid might a priori have been predicted on the basis that the bromination of phenol gives 90.2% para and 9.8% ortho derivative48~~~ and hence occurs some twenty times faster in the para position than in a single ortho position. More suitable for comparison with the dibromohydroxybenzoic acids are the relative rates of bromination of 2,6- and 2,4-dibromophenol. Such data as are available upon these phenols in Table V suggests that 2,6-dibromophenol is brominated more readily under fixed conditions than 2,4-dibromophenol but that the differences in reactivity are not as great as the phenol analogy suggests. Rediscussion of 2,6-Dibromophenol.-It is of interest to consider the possibility that 2,6-dibromophenol, ArOH, may undergo bromination according to a mechanism somewhat analogous to mechanism 111, namely, by the process (mechanism 111') fast ArOH & - ArO-
+ H + ; K: = [H+][ArO-]/[ArOH] (15')
ArOArOH
(44) Additional
+ Brz
+ Brz
k: ArBrO
k'-
ki ArBrO
k'-z
+ Br-
(16')
I
+ H + + Br-
(17')42
(45) See ref. (34), p p . 3 4 5 4 5 0 . (46) A. F. Holleman, Ckem. Reus., 1, 218 (1924). (47) For a possible explanation of this orlko-para ratio see P. W. Robertson, P. B. D. de la Mare and B . E. Swedlund, J . Ckem. Soc., 782
(1953).
577
BROMODECARBOXYLATION OF BROMOHYDROXYBENZOIC ACIDS
Feb. 5 , 1956 ArBrO
k! + Ar’Br-
+ H+
(19‘)4a
Ar‘Br- is the anion of tribromophenol and the reactive intermediate ArBrO is presumed to have the structure I11 analogous to that (11) from the 0
0
V
IV
I1
0
VI
VI1
I11
corresponding carboxylic acid. Such a mechanism leads to the kinetic consequence I/K*[H+] = k’-~[Br-]/k:kjKi
0
+ I / ( K:[H+l + k X i )
aration of the cyclohexadienone V from the bromination of 4-methyl-2,6-di-t-butylphenol in 80yo aceThis mechanism offers an alternative explanation tic acid. Bromination of 2,4-dibromo-3-methyl-6to that given earlier for the small decrease in k* t-butylphenol gives VI according to Forman and with increase in bromide ion concentration. If the SearPd; likewise chlorination of 2,4-di-t-butyl-5first term on the right-hand side of equation 27 is methylphenol yields VII. Analogous products of small compared t o the second right-hand term, then nitration have been described49and the reaction5” equation 27 reduces to equation l l a . Evaluation of 2,4,6-trialkylphenols with alkylperoxy radicals of the constants of equation 27 by a method similar gives products of related structure. The structure to that used for dibromosalicylic acid gives k’ of these products has been established especially on equal 3.2 I./mole sec., k:K’ 0.63 set.-', and k ’ - ~ / the basis of their ultraviolet and infrared absorpki 0.26 l./mole. The values of k: and k:K; are tion spectra. T h a t such 2,4,6-trisubstituted pheclose to the values of ko and k- given earlier (see nols undergo halogenation and nitration chiefly a t equation l l a ) and the reversibility ratio k ’ - l / k ; the position para to the phenolic hydroxyl group is is indeed so small that the first term on the right- in agreement with the fact that dibromo-p-hyhand side of equation 27 makes no more than a 17% droxybenzoic acid gives intermediate I1 considercontribution to the sum of the right-hand terms ably faster than dibromosalicylic acid gives I. a t the highest bromide ion concentration studied. Mechanism of Iodination of Phenols.-While we Thus in contrast to intermediate I1 from dibromo- have shown that the intermediate 111 from 2,6p-hydrobenzoic acid, intermediate 111 from 2,6-di- dibromophenol shows little if any tendency under bromophenol shows little, if indeed any, tendency our experimental conditions to revert back to the to revert to the starting phenol; this fact must re- starting phenol by reaction with halide ion, as sult chiefly because intermediate I11 loses a proton opposed to loss of a proton, i t occurs to us that an much faster than I1 loses carbon dioxide. Finally intermediate such as VI1 from the iodination of k:K: and ki for 2,6-dibromophenol are some ten phenol should show a greater tendency to do so. and thirty times larger than the corresponding con0 stants for dibromo-p-hydroxybenzoic acid; such is I1 t o be expected for an electrophilic substitution since the carboxyl group relative to hydrogen must lower the availability of electrons a t the atom to which i t is attached. i\ Reactive Intermediates in a Halogenation of H I VI1 Phenols.-We have proposed that reactive intermediates of structures I, I1 and I11 are formed This idea is supported by the greater nucleophilic during the bromination of the present phenols. characters1 of iodide ion versus bromide ion and by An alternative structure for the intermediate the probably greater ease of nucleophilic displacefrom the phenolic carboxylic acids is of type I V as mentS2upon iodine as opposed to bromine. Addirepresented for dibromo-p-hydroxybenzoic acid. tional evidence is supplied by the work of Soper We favor structures I and 11, however, since they and Smiths3 upon the kinetics of the iodination of give a simpler mechanism for the bromodecarboxyl- phenol. These workers found that for iodination ation. For structure I V t o give the final products with iodine in aqueous solution containing much rearrangement of bromine must occur before or iodide ion (0.1-0.4 M ) that the variation of the apsimultaneously with the decarboxylation step of parent second-order rate constant with iodide ion mechanism 111. Stronger evidence in favor of concentration was in agreement with a species such structures 1-111 is provided by the formation and (49) H.E.Albert and W. C. Sears, ibid., 76,4979 (1954). isolation of analogous, though more stable, interme(50) T.W.Campbell and G . M. Coppinger, ibid., 74, 1469 (1952); diates, from 2,4,6-trisubstituted phenols. Thus A. F. Bickel and E. C. Kooyman, J . Chcm. SOC.,3211 (1953). (51) C. G.Swain and C. B. Scott, THISJOURNAL.,75,141 (1953). Coppinger and Campbell@have described the prep(52) For some evidence see J. Hine and W. H. Brader. ibid., 77,361 (27)
0
(48) G. M. Coppinger and T. W. Campbell, THISJOURNAL, 75, 734 (1953).
(1955).
(53) F. G.Soper and G . F. Smith, J. Chew. SOL, 2757 (1927).
N. C. DENOAND RICHARD STEIN
578
as hypoiodous acid as the effective iodinating agent. X more complete studyS4gave the empirical result rate = k,’[PhOH][12]/JHil[I-] 4-
HA [PhOH] [Iz1 [HA]/[ H +I ‘[I-]
(28)
where HA is a buffer acid such as acetic acid. The first term was interpreted as involving attack of HOI upon un-ionized phenol or alternatively I + (or HzOI+) upon phenoxide ion while the second term involved attack of acyl hypohalite upon phenoxide ion (or a general acid-catalyzed attack of hypoiodous acid upon phenoxide ion). We wish to point out that equation 28 is just the kinetic form which would be expected for mechanism 111’ (involving iodine in place of bromine) with steps 16’ and 17’ essentially a t equilibrium and, therefore, with the second right-hand term of equation 27 small relative to the first. The first right-hand term of equation 28 would then result from water being the general base which accepted the proton of step 19’ while the second term would involve the base Aconjugate to the buffer acid. Moreover if mechanism 111’corresponds to the true mechanism of halogenation of phenol, then molecular iodine is seen to be the effective iodination agent. This iodinating agent has been shown to be responsible for the iodinolysis of p-methoxybenzeneboronic55 acid in aqueous solution a t iodide ion concentrations of 0.1 to 0.5 11.1 and is apparently responsible for the ex-
\Tal. 78
change reaction between radioactive iodine and diiodotyrosine.b6 At extremely low iodide concentrations such as are maintained by excess silver ion, the hypoiodous acidium ion (or iodinium ion) is probably the effective iodinating agents?; our question is whether in solution of high or moderately high iodide ion concentration the concentration of the hypoiodous acidium ion is sufficiently great to compete successfully with the much larger concentration of molecular iodine.58 A final answer to this question for phenol must await a study of the kinetics of iodination to much lower iodide ion concentration or, perhaps better, a study of a possible hydrogen isotope effect here as a function of the iodide ion concentration. If mechanism 111’can be shown to apply to the kinetics of the iodination of phenol,59 then probably a similar mechanism holds for the iodination of aniline.60 Acknowledgments.-The authors are indebted to the Graduate School for a fellowship to U. V. H. which made part of this work possible and to Dr. Jack Hine for helpful discussions.
(54) B . S. Painter and F. G . Soper, J . Chem SOL., 342 (1917); see also E . Berliner, THISJOURNAL, 73, 4 3 0 i (1951). ( 5 5 ) H. G. Kuivila and R . M. Williams, THISJOURNAL, 76, 2679
(56) A. H. Zeltmann and M. K a h n , ibid., 76,1554 (1954). (57) D. H. Derbyshire and W. A. Waters, 1. Chem. Soc., 3694 (1950); I. R . L. Barker and W. A. Waters, ibid., 150 (1952); R. P. Bell and E . Gelles, ibid., 2734 (1951). ( 5 8 ) F o r a similar and more detailed argument concerning t h e reaction of oxalic acid with bromine O C ~ S Y Shypobromous acidium ion see Y. Knoller and B. Perlmutter-Hayman, THISJOURNAL, 77,3212 (1955). (59) A somewhat similar mechanism has been considered for the iodination of 2,4-dichlorophenol b u t without appreciable experimental confirmation ( J . E. T a y l o r a n d M. I. Evans, Ohio J . Sci., 6 3 , 3 7 (1953)). (60) E. Berliner, THIS JOURNAL, 72, 4003 (1950).
(1954).
ATLANTA, GEORGIA
[COSTRIBUTIOS FROM THE COLLEGE OF CHEMISTRY AND PHYSlCS, PENNSYLVAXA STATE UNIVERSITY]
Carbonium Ions. 111. Aromatic Nitration and the Co Acidity Function1 BY X. C. DENOAND RICHARD STEIN RECEIVED SEPTEMBER 8, 1955 The rates of aromatic nitrations have been studied as a function of sulfuric acid Concentration.
The work reported in this paper is part of the third step in a series of investigations carried out in aqueous sulfuric acids whose ultimate goal is to determine whether or not aliphatic tertiary alkyl cations exist as reaction intermediates. The first step in the program was to evaluate an acidity function (Co), defined by eq. 1, where ~ K RisCthe negative logarithm of the equilibrium constant for eq. 2 .
+
Co = (PKR*) log ( C R O H / C R + ) H20 4- R’ = ROH H+
+
(1) (2)
The second step was to test the generality of eq. 1 with data from a variety of equilibria in aqueous sulfuric acid. These two steps were accomplished in the initial paper of the series2in which a number of triaryl, diary1 and monoaryl groups were used for “R.” ( 1 ) Grateful acknowledgment is made of the support of this research by a grant from the Petroleum Research Fund of the American Chemi. cal Society. (2) N. Derio, J J . Jaruzcrlski ;$nil A Schrieaheim. T r r r s Ti)irewi\r,, 7 7 , 8044 (1Q5:).
The third step in the program involved testing the generality of eq. 3 with kinetic data from appropriate acid-catalyzed reactions of the type of eq. 4 in which the g-roup R was varied as widely as possible. Equation 3 follows directly from reaction path 4 if eq. 5 is valid.2 In eq. 5 , R + on the righthand side of the equation represents the cations used to evaluate Co and R on the left-hand side refers to the same R as in reaction path 4. The SUperscript (*) in eq. 5 signifies the transition state for path 4, and “f“ is an activity coefficient. ROH
+ H+ =
log k = -Co H20
+ R+ L
+ constant
(3)
-k A -+- products ( 4 ) rate determining
d log (fAfROH/f*)d
% %SO4
=
% HzS04 ( 5 ) The fourth step can be undertaken if eq. 3 proves satisfactory for a wide variety of R groups, This; d log (fnoa/fR+)/d
